From: STEVE HARDING <Steve.Harding@nottingham.ac.uk>
  To  : Jeffrey A. Cohlberg <cohlberg@csulb.edu>
  Date: Sun, 6 Sep 1998 16:41:44 GMT0BST

Re: best program for sed eq analysis?

> Date:          Fri, 04 Sep 1998 16:34:34 -0700
> From:          "Jeffrey A. Cohlberg" <cohlberg@csulb.edu>
> Organization:  Cal. State Univ. Long Beach
> To:            rasmb@bbri.harvard.edu
> Subject:       best program for sed eq analysis?

> Dear RASMBers,
> 
> I'm a Model E user, with the output voltages that drive the scanner 
> recorder being fed to a Macintosh computer via an analog-to-digital 
> converter controlled by home-made software.  Which of the computer 
> programs available for downloading (NONLIN, MSTAR, others?) would be 
> best for me to use to analyze data from routine sedimentation 
> equilibrium experiments (getting molecular weights and assessing 
> heterogeneity, not necessarily doing rigorous analysis including 
> determining Ka's)?  I prefer Macintosh but can use PC if necessary.
> 
> Thanks for your feedback.
> 
>      Jeff Cohlberg
> -- 
> Jeffrey A. Cohlberg, Professor
> Department of Chemistry and Biochemistry
> California State University, Long Beach
> 1250 Bellflower Blvd.
> Long Beach, CA 90840
> phone (562) 985-4944	fax (562) 985-8557
> 

Jeff:
We use Model E's as well, but only for Rayleigh and Schlieren work. 
Traditionally we used an MSE Centriscan (we still have one) for doing 
the uv work but that has largely been taken over by the XL-A and 
XL-I.
MSTAR (for a PC, not MAC) is set up for receiving absorbance data 
from the XLA and XLI (MSTARA)  and also Rayleigh interference data 
from the XL-I and Model E (MSTARI), but not scanner data from the E.  
However it should be trivial to change a couple lines in the 
programme to receive your data.  
If you'd like to use MSTARA probably the best thing is if you email 
me a sample data file and we can do the modifications here.  For data 
processing you need the rotor speed, vbar, solvent density  
temp, radial positions of the meniscus and base, and also the 
baseline absorbance (we try to avoid floating baselines). MSTAR seems 
to do what you want: a model independent routine for (i) giving the 
weight average mol. wt over the whole distribution (from meniscus to 
base, not just where the data truncates) (ii) ln(conc) versus r^2 and 
point weight average plots (versus r^2 and conc - without 
over-smoothing) to give an idea of heterogeneity (uv absorbance data 
tend to give rather noisy point average plots unless you smooth - 
hence the attraction of Rayleigh data.  MSTAR also gives you the 
option of pulling out point z-average data and  but this is usually a 
waste of time with uv-absorption data because of the noise).  Having 
seen whats going on, then we use the more detailed model dependent 
routines such as PSI, NONLIN, or Beckman's ASSOC4 programme. 

Hope this helps  
 
Steve